How to use package sbim: Simulation-Based Inference using a Metamodel for Log-Likelihood Estimator2 years ago
Introduction | Mathematical framework | Parameter inference using local asymptotic normality (LAN) for simulation log-likelihood | Getting started | An example: gamma-Poisson observations | Hypothesis test on the maximum expected simulation log-likelihood estimator (MESLE) | Hypothesis test on the simulation-based proxy | Inference for non-iid observations | Stochastic volatility model | More on hypothesis tests | Test on the metamodel parameters | Test on the mean and variance of the simulation log likelihood at a single parameter value | Automatic weight adjustments for bias reduction | Proposal of the next simulation point for near-optimal efficiency | References